Objective: Diabetic retinopathy (DR) remains a leading cause of visual impairment. A frequent manifestation of DR is diabetic macular edema (DME) for which laser photocoagulation has been the only proven treatment for the last several decades. Studies have shown that anti-vascular endothelial growth factor (VEGF) injections such as bevacizumab or ranibizumab have been efficacious in treating patients with DME. However, there has been no direct comparison of these agents to determine whether one treatment is more effective than the other. The objective of this study is to compare the treatment efficacy of ranibizumab versus bevacizumab in eyes with DME. Study Population: Sixty (60) participants with macular edema secondary to diabetes and any stage of DR (other than those requiring scatter laser photocoagulation for proliferative DR) in one or both eyes will be enrolled in this randomized study. Design: In this Phase II, comparative, double-masked study, eyes will be randomly assigned to receive ranibizumab or bevacizumab. During the initial phase of the study participants will participate in a three-period, 36-week, cross-over study in which study eyes will be assigned to one of four treatment groups (i.e., treatment sequences). The two drugs and three periods form an AAB/ABB/BBA/BAA pattern as follows: Group 1 eyes will receive a series of intravitreal injections of ranibizumab at baseline and Weeks 4, 8, 12, 16 and 20, then cross-over to receive a series of intravitreal injections of bevacizumab at Weeks 24, 28 and 32. Group 2 eyes will receive a series of intravitreal injections of ranibizumab at baseline and Weeks 4 and 8, then cross-over to receive a series of intravitreal injections of bevacizumab at Weeks 12, 16, 20, 24, 28 and 32. Group 3 eyes will receive a series of intravitreal injections of bevacizumab at baseline and Weeks 4, 8, 12, 16 and 20, then cross-over to receive a series of intravitreal injections of ranibizumab at Weeks 24, 28 and 32. Group 4 eyes will receive a series of intravitreal injections of bevacizumab at baseline and Weeks 4 and 8, then cross-over to receive a series of intravitreal injections of ranibizumab at Weeks 12, 16, 20, 24, 28 and 32. Participants for whom one eye is enrolled in the study will have this eye randomized to one of the four groups above. Participants for whom both eyes are enrolled in the study will have the right eye randomized to one of the four groups above. For those whose right eye is randomized to Group 1 or 2, the left eye will be randomized to Group 3 or 4; conversely, for those whose right eye is randomized to Group 3 or 4, the left eye will be randomized to Group 1 or 2. Following this cross-over phase, eyes will be returned to the treatment (ranibizumab or bevacizumab) to which they were originally assigned and treated on an as-needed basis through a common termination date three years from enrollment of the last-enrolled participant. Both the treating investigators and participants will be masked to the group assignments. The primary outcome will be assessed at Weeks 12, 24, 36 and at Years 1, 2 and 3. Outcome Measures: The primary outcome measure is the mean change in best-corrected visual acuity (BCVA). Changes in BCVA from baseline to four weeks following the end of each of the three periods (i.e., Weeks 12, 24 and 36) and at Years 1, 2 and 3 will be used for the primary analysis. Secondary outcomes (assessed between the baseline and Week 12 visits, Weeks 12 and 24 visits and Weeks 24 and 36 visits and at Years 1, 2 and 3) will include the mean changes in central macular thickness and central retinal volume by treatment group as measured by OCT; the slope of the changes in BCVA, central macular thickness and retinal volume; the proportion of eyes with visual improvement greater than or equal to 10 letters; the proportion of eyes with visual improvement greater than or equal to 15 letters; the proportion of eyes with greater than or equal to 0.1 log unit loss or gain in logOCT; the proportion of eyes with greater than or equal to 0.05 log unit loss or gain in logOCT; changes in fluid leakage in the macula as demonstrated by fluorescein angiography; and changes in macular structural improvement (i.e., resolution of cystic changes) as measured by OCT. Other secondary outcomes will include the proportion of eyes meeting criteria for significant worsening, treatment success, or treatment failure, the frequency of re-injection among eyes in the treatment-as-needed phase of the study, and the proportion of eyes receiving focal/grid laser photocoagulation or other adjuvant treatment during the course of the study. Safety outcomes include the number and severity of adverse events. The number of eyes withdrawn from the investigational product due to vision loss or adverse events and the number of eyes deemed to have worsening disease will also contribute to the assessment of safety. Fifty-six (56) participants were enrolled in the study and enrollment was closed. Fifty-five (55) participants completed the 36-week crossover phase of the trial in fiscal year 2014. Results of the 36-week crossover study were presented at the Association for Research in Vision and Ophthalmology (ARVO) 2015 Annual Meeting (Denver, CO) in May 2015. Results of the 36-week crossover study have been submitted for publication. The last participant close-out visits for the extension phase of the study were completed in early 2015, and the trial was terminated.